BlinkBook
Knowledge Source had the idea, and we founded the company that became Cloud37 to build it. I built the demo that started it all, and at launch blinkbook.ai earned over $1M in its first two weeks.
available for senior AI engineering roles
Patrick Ward
Software engineer who ships production LLM systems end to end: retrieval, agents, tools, billing, Kubernetes and observability.
- blinkbook.revenue[0:14d]
- $1M+
- models.in_production
- 45+
- model_spend / month
- $10k+
- cloud37.commits
- ~1,100
// what I've built
Cloud37 platform
Multi-tenant SaaS where businesses build, test and sell AI agents. I bootstrapped the codebase in 2024: 270,000 lines across FastAPI and Next.js.
Tables: safe SQL for agents
Model-written SQL parsed against an allowlist, tenant-scoped, and run as a read-only Postgres role with timeouts and row caps. Backed by 323 tests, security tests included.
Hybrid retrieval
Qdrant and Elasticsearch combined with rank fusion, query expansion and reranking, with graceful fallbacks that keep answers flowing.
25 connectors
The framework that gives agents tools in Gmail, Calendar, Drive, Slack, Notion, GitHub, Outlook, Stripe, QuickBooks and more, metered per call.
Cost accounting
Token accounting across input, output, cached and reasoning tokens for 45+ models, reconciled against OpenAI, Anthropic and Google bills.
// git log --author="Patrick Ward"
- 2024 to nowCloud37Software Engineer, AI Specialist
- 2021 to 2024GrowRev DigitalTech Engineer, Media Buyer
- 2016 to 2021WorldWide Tech ServicesService Technician
- 2006 to 2008Tunxis Community CollegeComputer Programming